BACKGROUND: Vaginal intraepithelial neoplasia (VAIN) is a rare precancerous lesion, and early diagnosis is crucial for preventing its progression to invasive vaginal cancer. However, the subtle differences in morphology and color between VAIN lesions and normal vaginal tissue make the automatic segmentation of VAIN highly challenging. Existing methods struggle to achieve precise segmentation, impacting the efficiency of early screening.; PURPOSE: This study aims to develop a high-accuracy, robust deep learning image segmentation network to accurately and automatically segment VAIN lesions, thereby improving the efficiency and accuracy of early VAIN screening.; METHODS: We propose a multi-scale dilated attention flow network for VAIN image segmentation. This network improves upon the U-Net architecture by optimizing the designs of the encoder and decoder and incorporating skip connection modules. In the encoding stage, we introduce the dilated squeeze-and-excitation (DiSE) module and the flow field guided adaptive separation and enhancement (FGASE) module. The DiSE module integrates dilated convolutions with varying dilation rates and a channel attention mechanism, effectively extracting multi-scale contextual information and enhancing the model's ability to perceive VAIN lesions of different sizes. The FGASE module employs flow-guided techniques to dynamically separate the features of the main region (VAIN lesions) from the edge region and enhance them individually. In the decoding stage, we propose a depth wise enhanced pooling (DEP) module that combines deep convolutional layers with adaptive pooling strategies to improve local feature extraction capabilities and optimize global contextual information. The skip connection stage introduces a triple statistical attention (TSA) module that utilizes global average pooling, global max pooling, and global standard deviation pooling to effectively capture diverse feature information, thereby enhancing the model's ability to model long-range dependencies.; RESULTS: Experiments conducted on a VAIN image dataset comprising 1142 patients demonstrate that the proposed network significantly outperforms other medical image segmentation methods across six metrics: Mean intersection over union (MIoU), dice coefficient, accuracy, recall, precision, and mean absolute error (MAE). Specifically, this network achieved an MIoU of 0.8461 and a Dice coefficient of 0.9166, substantially higher than other comparative methods, with a faster convergence speed. Ablation studies further confirm the effectiveness of each module in enhancing the model's performance.; CONCLUSIONS: The proposed network exhibits exceptional performance and robustness in the task of VAIN image segmentation, effectively segmenting VAIN lesions and providing strong technical support for early VAIN screening and clinical diagnosis. This work has significant clinical application value. © 2025 American Association of Physicists in Medicine.

从信标光链路中高效地获取运动平台的姿态信息是运动平台量子光通信成功的关键之一。针对运动平台量子光通信的实际需求,提出一种探测载体平台旋转姿态角的信标姿态解耦方法,有望显著提高量子光通信的动态链接效率。经0°、45°

为在小腔体条件下拍摄人体内部表面血管组织的图像，准确地将血管组织状态呈现出来，改善常规图像探测器尺寸大、不均匀光照条件下成像质量差、血管成像对比度低等问题，设计基于OV6946超小型图像探测器、具有300nm-1100nm宽谱成像能力、国产盘古FPGA驱动的电子内窥图像采集系统，研究改进限制对比度自适应直方图均衡化图像增强算法(Contrast Limited Adaptive Histogram Equalization, CLAHE)，实现人体小腔体组织表面的血管图像增强功能。通过对比验证，本系统在非均匀光照下，针对小腔体内部组织成像具有更好的成像效果，改进CLAHE图像增强算法在提升血管组织成像对比度上具有更强的优势，为医用超细电子内窥系统或有类似需求的宽谱图像增强系统的推广，提供了一种有效的技术解决方案。

Aiming at the interferences caused by vibrations on the target monitoring system of satellite platforms, an experimental system of optical-mechanical-electrical joint image stabilization based on inertial sensors and piezoelectric fast mirrors was studied. First of all, the data from the fiber optic gyroscope and the detector are combined and processed. After that, the predicted intentional velocity is subtracted from the real-time gyroscope data, and the vibration displacement is obtained using an integral. The feedback amount is calculated using the processor according to the vibration displacement, the piezoelectric fast mirror is controlled to compensate the vibration interferences, and the purpose of stabilizing the images is achieved. The experimental results show that under a vibration interference of 300 μ rad /s, the image stabilization control frequency of the system reaches 500 Hz, the relative error between image frames is better than 3 μ rad , and the intra-frame stabilization performance of the detector meets the requirements on stabilized image and target detection for space targets on the star-carrying platform. © 2024 Author(s).

Bound states in the continuum (BICs) with infinite quality factor (Q-factor) and significant field enhancement pave the way for realizing highly sensitive optical sensors with enhanced light-matter interactions on the nanoscale. However, current optical sensing methods are difficult to discriminate between isotropic and anisotropic media from resonance spectral lines, resulting in optical sensing methods still being limited to isotropic media. In this work, we demonstrate that BICs can be realized by modulating the period of structural units to convert BICs to QBICs without changing their space group symmetry, and propose a polarization-independent metasurfaces-based realization of highly sensitive refractive index sensors for isotropic and anisotropic media as well as discrimination. We propose a metasurface of tetrameric silicon nanoboxes with C4 symmetry as structural units to achieve the conversion of BICs to QBICs by modulating the period of structural units without changing the geometry of the structure. Two QBICs modes dominated by electric toroidal dipole and magnetic toroidal dipole are identified by multipolar decomposition and electromagnetic distribution calculations. Meanwhile, we realize the refractive index detection and resolution of isotropic and anisotropic media based on polarization-independent metasurfaces combined with isotropic and anisotropic media layers. Our work provides what we believe to be a new method for realizing the fast resolution and refractive index optical sensing of isotropic and anisotropic media, and offers new ideas for the design and application of polarization-independent metasurfaces.

Circular dichroism (CD) effects have broad applications in fields including biophysical analysis, analytical chemistry, nanoscale imaging, and nanosensor design. Herein, a novel design of a tilted F-type chiral metal nanostructure composed of circular nanoholes with varying radii has been proposed to achieve remarkable CD effects, and the results demonstrate the generation of a significant current oscillation at the sharp edges where the nanoholes overlap under circularly polarized light, resulting in a strong CD effect. The CD effect can reach up to 7.5%. Furthermore, spectral modulation of the resonant wavelength can be achieved by adjusting the structural parameters, which enhances the tunability of the structure. Overall, these results provide theoretical or practical guidance for enhancing the circular dichroism signal strength of chiral metal nanostructures and designing new types of two-dimensional chiral structures.

For a long time, wind speed profile measurement has been the primary task of weather forecasting. Therefore, the detection of atmospheric wind speed is extremely important for studying the changes in atmospheric motion. In order to solve the problems of insufficient data collection, low resolution, and low accuracy in atmospheric wind field detection, this paper introduces the relevant theories of wind speed detection, completes the optical design of the system according to the research objectives, and determines the selection of optical devices. At the same time, a Doppler wind lidar system based on a quadrichannel Mach-Zehnder interferometer is designed and built to carry out ground-based observation experiments, collect echo signal data, and inverse the atmospheric radial wind speed. Furthermore, the wind measurement error is analyzed. Firstly, the paper introduces the basic principle of the wind measurement system, i.e., using the Doppler effect of light, and then analyzes the frequency discrimination device of the system in detail, and obtains the theoretical calculation method of atmospheric wind speed inversion. At the same time, the relevant datasets of wind measurement system are analyzed, including backscattering ratio, aerosol, and molecular extinction coefficient, and the emission mechanism of the large pulse laser is also studied in detail, which provides a theoretical basis for the model construction of Doppler lidar and the research on the enhancement of pulsed laser emission energy. Secondly, according to the research index of wind measurement, a Doppler wind measurement lidar system based on a quadrichannel Mach-Zehnder interferometer is designed, including the design of ab external light path transceiver system, internal light path interferometer, software and hardware, and algorithm. The calibration of the quadrichannel Mach-Zehnder interferometer is completed, with its maximum interference contrast reaching 0.869. Through the self-developed optical transceiver system and data acquisition system, the echo signal of lidar is received and detected. Lastly, the data of echo signals collected by the interferometer are analyzed, the radial atmospheric wind speed profile is inversed, and the signal-to-noise ratio and wind speed measurement error of the system are evaluated. The experimental results show that the maximum signal-to-noise ratio (SNR) of the system can reach 1433 when the emission pulse energy of the large pulse laser is adjusted to 255 mJ, and the farthest wind speed detection distance is about 8 km. The high-precision wind speed detection range can reach 2 km, the actual wind measurement errors in this range are all within 1.593 m/s, and the minimum error is only 0.418 m/s. In addition, the backscattering coefficient and extinction coefficient of atmospheric molecules and aerosols in the range of 8 km and the atmospheric temperature in the range of 10 km are also measured. The measurement accuracy of the aerosol extinction coefficient is & PLUSMN;0.001 m(-1), and the measurement error of atmospheric temperature within 10 km is within 2 K, achieving the expected goals.

为了提高泊松表面重建算法效率并改善重建结果细节表现，采用一种基于混合树的点云搜索方法，平衡了八叉树和二叉树技术关于时间复杂度和空间复杂度的冲突。并在点云搜索阶段通过引入多个能量项对点云进行密度评估与滤波等，针对点云稀疏部分进行自适应的点云稠密化以保证重建模型的细节与准确度。实验结果表明，混合树重建算法与泊松表面重建算法和屏蔽泊松算法相比速度分别平均提升33%和15%，且能更好的保持重建模型的细节，误差最小。该研究为点云的表面重建提供了参考。

In the field of civil aviation safety and air security, research on identification and tracking of small aircraft is very important. Aiming at the problem of low recognition rate of long-distance small air targets, this study proposes a method of GST-YOLOv3 combined with DeepSORT algorithm to detect and track small aircraft. The proposed GST-YOLOv3 uses a ghost module to extract effective features and simplify the network. In feature extraction layer, the spatial pyramid pooling layer is added to fuse local features and global features to enhance the features of small targets. The Transformer module is embedded in the feature fusion layer to focus on global context information. In the multiscale detection layer, a new large-scale feature map detection layer is formed by fusing shallow features and deep features to obtain a smaller receptive field. To avoid frame loss of video detection, the DeepSORT algorithm is used for real-time tracking. In this study, the self-made data set Aplanes and the public data set VisDroneDET are used to verify the effect of the proposed GST-YOLOv3. The computation of the GST-YOLOv3 was reduced by 45.8% compared with YOLOv3; the missed detection rate of airplanes decreased by 2.5%; and the detection precision, recall rate, F1 value, and average precision/0.5 value of small aircraft reached 98.4, 98.7, 98.5, and 98.6%, respectively, which are better than those of similar literature. © 2023, AIAA International. All rights reserved.

In this paper, a dual position-sensitive detector-based vision measurement system camera is built instead of a traditional CCD camera. The 3D position information for the light point is calculated according to the 2D coordinate information of a certain light point in the space illuminated on the two position-sensitive detector (PSD) photosensitive surfaces, which is used for position detection of the spatial light point. In addition, the positioning model for 2D PSDs with different spot sizes in the Gaussian spot mode is derived by the mathematical model of Lucovsky’s differential equation for a PSD. For the nonlinear distortion of the PSD, a nonlinear error calibration method using a particle swarm combined with a back propagation neural network is proposed to correct the errors in the measured values through the relationship between the input and output values, to obtain the predicted value that approximates the real coordinates. Then, by comparing the influence of different spot sizes on the positioning accuracy, we conclude that the smaller the spot formed by the convergence of the beam under the optical lens, the higher the positioning accuracy. We believe this conclusion can help improve the accuracy of PSD measurements. Finally, a red LED light spot is set up, and the 3D position measurement and error calibration of the light spot is done by dual PSD cameras, which better solves the position detection problem of a space light spot under close-range conditions because it is fast, reliable, and easy to implement. It also provides an effective method to detect the motion trajectory of a moving light spot in space. © 2023 Optica Publishing Group.